Installed software was just called “software” until the advent of Software-as-a-Service (SaaS), the top tier in the Cloud infrastructure architecture. With the addition of this browser-based delivery mechanism, new applications and new business models emerged.
As exciting as SaaS is, the vast majority of software used in the world today is still installed on the device on which it is used. These applications range from the Microsoft Office suite for a few hundred dollars to advanced simulators for everything from semiconductor design and oil exploration that cost hundreds of thousands and sometimes over one million dollars per license.
With all the benefits that SaaS provides, some would think that traditional installed software would disappear. In this article, I outline the reasons it won’t, from two perspectives: why software vendors will continue to invest in developing installed software, and why software buyers will continue to buy it.
Why Software Vendors Still Invest in Installed Software
Despite the predictions by technology prognosticators that installed software will disappear and be replaced by SaaS applications, investment in traditional software development continues apace. The primary reason for this is business model inertia.
SaaS may have higher growth rates, but the installed software industry will remain several times bigger in revenue and market share for several years to come. Revenue estimates vary, but recent reports from PwC and Gartner place the non-SaaS enterprise software market between $130 billion and $150 billion in 2015. And growth rates aren’t bad, either. In another report released by Gartner in June 2011, they forecast that by 2015 worldwide enterprise application software revenues would grow by 12 percent. With sales and growth figures like these, software vendors will fight to protect their traditional software businesses.
One reason: it’s hard to abandon the pricing model. As consumers, we buy a one-time license to use installed software and then typically don’t make another purchase from that vendor until a new version of the software is released a few years later. In contrast, businesses purchase annual licenses to use software and pay the vendor an additional 15-20 percent of the license fee for access to software updates and technical support. This is a lucrative, high-margin business.
The market capitalization, share price, and therefore returns to investors in all of the companies that generate these billions of dollars at high margins are dependent upon their existing revenue streams and business models. This alone makes a transition to SaaS easy for vendors to resist.
On the product side of the equation, software vendors have spent decades developing, testing and refining millions of lines of code. The drive to entice a customer to upgrade to the next release is invariably through incremental development rather than deep innovation. Re-writing code under a completely different architecture, using different languages, programming models and user interface philosophies is often perceived as too large an investment and extremely high risk.
Even if the decision were made to re-write their products for SaaS, the vendor would have to either re-train all their developers or hire a new team. If you’ve tried to hire a competent SaaS user experience architect or developer recently, you’ll know how difficult a proposition this is.
An additional complication that a move to SaaS presents is building the Cloud infrastructure to run a new multi-tenant SaaS application. As Infrastructure-as-a-Service (IaaS) vendors continue to compete and differentiate their offerings, this aspect of the challenge should diminish over time. However, as with any business, vendors must identify and invest in their core competency. To stray away from this is expensive and risky.
Why Software Users Still Purchase Installed Software
We’ve covered the supply side. What about the demand side of the software business equation? If SaaS is so compelling, why are enterprises still buying installed software? Five reasons jump out.
- Security. This is the most frequently cited reason for sticking with installed software. Corporations don’t want their intellectual property (IP) outside of their corporate firewall and potentially “mingling” with their competitors’ IP on a multi-tenant SaaS platform. As the recent migration to Google by the Spanish bank BBVA shows, this is a barrier that is starting to erode, but it’s still very real for many corporations.
- The business model. Counter-intuitively, one of the biggest inhibitors to moving to SaaS is the business model. While switching from a capital expense to an operational expense can be attractive, the inherent unpredictability of SaaS application usage complicates the budget process for many enterprise purchasing departments. While not a huge impediment over time, it weakens one of the vaunted strengths of the SaaS model.
- The need for speed. Processing speed is a legitimate reason for not moving to SaaS for many high-end software products. Applications that perform graphics rendering and manipulation for massive computational problems such as simulation (DNA folding, fluid dynamics, aerodynamics, circuit design, etc.) need the performance of compiled software. The interpretive languages of the Web and SaaS are simply too slow to deliver the performance users require.
- The ecosystem. Big-ticket software applications rarely operate in isolation. Whether used for movie production, pharmaceutical development or chip design, software applications are invariably “modules” in a large and complex flow of software applications from different vendors. The resultant whole is much greater than the sum of its parts.
- Customization. Large consumers of software require from their vendors specific features or interface interoperability between different applications. This is often harder to deliver with SaaS than with installed software.
How Installed Software Can Leverage the Cloud
The issue of installed versus Cloud software isn’t completely exclusive. Installed software can use resources in the Cloud, too, in mostly the same way it would use a traditional data center. The key difference is that the virtualized resources of a Public Cloud result in much more efficient hardware utilization.
High utilization is particularly helpful for software used in High Performance Computing (HPC) environments. In HPC, a massive number of computations are executed across multiple machines in parallel, instead of sequentially on a single machine. Using HPC can therefore deliver results in hours or days for what used to take weeks or even months.
A Public Cloud can be seen as a source of effectively infinite compute resources available on short notice, thus enabling high scalability and flexibility. And because the underlying IT infrastructure is managed by the Cloud provider, this helps lower costs considerably. The combination of installed software with the infinite computational resources of the Cloud can provide a “best of both worlds” solution.
There are other combination installed software/Cloud opportunities, too. For instance, some vendors offer to install customers’ software in their Private Cloud. This model can deliver high-value services with the software, a proof-of-concept experiment, or as a way to sell the customer more licenses when they don’t have the in-house resources to run them. In addition, “Hybrid Clouds” offers an independent environment where end-users can more easily work with software from multiple vendors without violating any licensing constraints.
Installed software isn’t going away anytime soon. Software vendors will fight to stick with their high-revenue, high-margin businesses, and corporate customers still see a variety of benefits from it. However, new Cloud models are emerging that enable some crossover between installed software and the Cloud, and business models will continue to evolve in all arenas. Things are just getting started.